Vitamin K2 for Type 2 Diabetes: Promising, Not Proven

Vitamin K2 is not a diabetes treatment, but a six-month randomized trial gives it a more credible signal than most supplements marketed for blood sugar. In 60 adults with type 2 diabetes, 90 micrograms per day of menaquinone-7 (MK-7) lowered fasting glucose, fasting insulin, and HbA1c compared with placebo yogurt. The same research group also transferred stool from supplemented mice into other mice and saw better glucose tolerance in the recipients.

That mouse experiment makes an interesting claim possible: changes in gut microbes may help carry part of MK-7's metabolic effect. It does not prove that claim. The human trial was small, the supplement came in yogurt rather than a neutral capsule, and the fecal-transplant design cannot rule out direct effects from residual vitamin K2 in donor stool. Still, the clinical result deserves more attention than a jar-label promise.

What the main trial found

In a 2023 double-blind, placebo-controlled trial by Yuntao Zhang and colleagues, 60 community-recruited adults with type 2 diabetes received either 90 micrograms of MK-7 daily in 100 grams of pasteurized yogurt or identical yogurt without MK-7. The trial lasted six months, with 30 participants in each group.

After adjustment for baseline values, the MK-7 group had a 13.4% reduction in fasting plasma glucose, a 28.3% reduction in fasting insulin, and a 7.4% reduction in HbA1c relative to the placebo group. Each comparison reached statistical significance at P<0.05. Measures derived from the HOMA2 model also indicated improved insulin sensitivity.

Those are meaningful changes if larger trials reproduce them. HbA1c and fasting glucose are clinically relevant markers, but the study did not test whether MK-7 prevented diabetic complications, reduced medication needs, or improved long-term health. Nor should percentage changes be mistaken for an individual prescription: the publication's small sample means the reported effect sizes have substantial uncertainty.

The microbiome result is intriguing, not settled

The investigators measured participants' stool bacteria with 16S rRNA sequencing and measured selected fecal metabolites. Compared with placebo, the MK-7 group showed more Bacteroides and other putative short-chain-fatty-acid-producing taxa, along with higher butyrate and lithocholic acid. The authors proposed a chain in which microbial changes increase short-chain fatty acids and secondary bile acids, influence TGR5 and FXR signaling in the colon, increase GLP-1 secretion, and improve insulin sensitivity.

Every link in that chain needs independent testing. A change in bacterial abundance alongside a better HbA1c does not establish that the bacteria caused the change. It could be part of the pathway, a response to another metabolic shift, or a marker that travels with the real mechanism.

The mouse work pushed beyond correlation. Stool from high-fat-diet mice given MK-7 was transferred into antibiotic-cleared mice also fed a high-fat diet. At week 16, recipients of stool from supplemented donors had improved oral glucose tolerance. This is a preclinical finding, not a human fecal-transplant trial, but it is an attempt to test whether a transferable gut factor contributes to the effect.

There is a large catch. The donor mice were receiving MK-7 when researchers collected their stool. That material may have contained residual vitamin K2 as well as altered microbes. Recipients could therefore have received a direct pharmacologic exposure rather than a microbiome-only intervention. Germ-free recipients given defined microbial communities, or donor washout before stool collection, would offer a cleaner answer.

Why the hypothesis remains biologically plausible

The broader idea that gut-directed interventions can affect diabetes markers has support. A 2024 systematic review and meta-analysis of prebiotics and Mediterranean and plant-based diets reported improvements in gut microbiota measures and glucose control among people with prediabetes or type 2 diabetes. A meta-analysis of higher microbiota-accessible carbohydrate intake also evaluated cardiometabolic effects in type 2 diabetes, where microbial short-chain-fatty-acid production is a plausible route.

More targeted interventions have produced similar signals. A meta-analysis of randomized probiotic trials assessed glycemic homeostasis in adults with type 2 diabetes. In a separate 12-week, double-blind trial of 120 people with type 2 diabetes, a synbiotic combining Bifidobacterium animalis MN-Gup with galactooligosaccharide outperformed the probiotic alone for glycemic control; see the 2025 trial. These studies support the general proposition that changing the gut environment can influence metabolic markers. They do not establish MK-7 as a prebiotic or validate Zhang and colleagues' specific pathway.

Vitamin K2 may also work without the microbiome. A 2025 skeletal-muscle study reported that vitamin K2 alleviated insulin resistance associated with muscle atrophy through AKT/mTOR signaling. Animal and cellular work has also implicated osteocalcin carboxylation and GLP-1-related pathways. Multiple plausible mechanisms are useful for hypothesis building, but they make it harder to assign the human trial's benefit to any one pathway.

Does another human trial support it?

Yes, directionally. A randomized controlled trial by Rahimi Sakak and colleagues in people with type 2 diabetes reported improved glycemic control with vitamin K2 supplementation. That matters because it came from a separate trial in the same disease population. The available report does not provide enough recoverable detail here to compare its effect sizes directly with the 2023 study.

Reviews put the topic in context but do not erase the uncertainty. A 2023 systematic review and meta-analysis of vitamin K and type 2 diabetes synthesized intervention and observational evidence, while a 2022 network meta-analysis compared nutritional supplements for glycemic control. The accessible reporting does not provide vitamin-K2-specific pooled estimates needed to judge whether the 13% to 28% changes in the small Zhang trial are typical or unusually large.

The null trial matters

MK-7 did not improve insulin sensitivity in a randomized trial of postmenopausal women, despite increasing plasma adiponectin. The Rønn et al. trial is not directly comparable with trials in adults who already have type 2 diabetes: the population and metabolic baseline differed. But it rules out the idea that MK-7 reliably improves insulin sensitivity in everyone.

People with established type 2 diabetes may have more room for metabolic improvement, or a different baseline microbiome, than metabolically healthier participants. That is plausible, not demonstrated. Geography, diet, genetics, and background vitamin K intake can all affect microbiome composition and may affect response. The main positive trial came from one site, and it did not report dietary vitamin K assessment.

Should people with type 2 diabetes take MK-7?

MK-7 at 90 micrograms per day is reasonable to discuss with a clinician as an adjunct, especially for someone with type 2 diabetes who wants a low-dose nutritional intervention with two directionally positive randomized trials behind it. It should not replace glucose-lowering medication, dietary treatment, exercise, or routine monitoring. Anyone using warfarin or another vitamin-K-sensitive anticoagulant should not start vitamin K supplementation without medical guidance.

The fairest verdict is moderate confidence in a glycemic signal for MK-7 in established type 2 diabetes and low confidence in the claimed microbiome mechanism. The human results need larger, independent trials with a non-yogurt delivery vehicle, measured background vitamin K intake, and enough participants to estimate the true size of the effect. Until then, the mouse transplant result is a worthwhile lead, not a finished explanation.